Combining cardiac and renal biomarkers to establish a clinical early prediction model for Cardiac surgery-associated acute kidney injury: a prospective observational study

Abstract Background Cardiac surgery-associated acute kidney injury (CS-AKI) is common and associated with poor outcomes. Early prediction of CS-AKI remains challenging. Currently available biomarkers for AKI are serum cystatin C (sCysC) and urinary N-acetyl-β-D-glucosaminidase (uNAG), and common cardiac biomarkers are N-terminal pro B-type natriuretic peptide (NT-proBNP) and cardiac troponin I (cTNI). This study aimed to evaluate the efficacy of these biomarkers in predicting CS-AKI. Methods Adult patients after cardiac surgery were included in this prospective observational study. The clinical prediction model of CS-AKI was established by the least absolute shrinkage and selection operator (LASSO) regression, and the discriminative ability of the model was evaluated by using the area under the curve of the receiver operating characteristic (AUC-ROC), decision curve analysis (DCA), and calibration curves. The risk nomogram was plotted, and the validation cohort was constructed for external validation. Results In the modeling cohort of 689 and the validation cohort of 313, the incidence of CS-AKI was 29.2% and 39.6%, respectively. Predictors screened by LASSO included age, history of hypertension, baseline serum creatinine, coronary artery bypass grafting combined with valve surgery, cardiopulmonary bypass duration, preoperative albumin, hemoglobin, postoperative NT-proBNP, cTNI, sCysC, and uNAG. The ROC-AUC of the constructed clinical prediction model in the modeling cohort and validation cohort were 0.830 (0.800–0.860) and 0.840 (0.790–0.880), respectively, and the calibration and DCA showed good fit and clinical benefit. Conclusions A clinical early prediction model consisting of the immediately postoperative renal biomarkers sCysC and uNAG and the cardiac biomarkers NT-proBNP and cTNI could improve the predictive accuracy of CS-AKI.